Photovoltaic power generation prediction method based on improved long short-term memory neural network

• Main Authors: GOU LINFENG, FENG GUANXIANG, CHEN HUATAO, URAD IVANOV, WU ZHIHAN, CHEN YINGXUE
• Format: Patent
• Language: Chinese; English
• Published: 18.07.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to a photovoltaic power generation prediction method based on an improved long short-term memory neural network, and the method employs an EBS optimization algorithm proposed by Mohsen Shahrouzi, further proposes an OLSTM model after providing the hyper-parameters of an LSTM neural network composed of original learning data and an LSTM layer, and will obtain an optimization result with higher prediction precision and real-time performance than a conventional model. The LSTM neural network is composed of m hidden layers with m * n units and a dense layer, high-precision hyper-parameter selection can be completed by using an EBS algorithm, a hyper-parameter XGbest is input by using an OLSTM model, normalized data is trained through an LSTM training function and a prediction function in an MATLAB toolbox, then power is predicted, and finally a prediction value is obtained. Compared with a traditional LSTM model, the OLSTM based on the combination of EBS optimization and the LSTM neural netw